SINAD Measurements

[signalpath]

SINAD data is the among the best data available to select well designed products. I don't see the need to separate THD from noise since less is better for both but there is no harm in providing both.- Rich

SINAD is great! Valid measurement. The problem is ASR's calling it a definitive measure of "guaranteed transparency." That's not science. That's belief.

 

[pma]

SINAD as used here at ASR for amplifiers is a measurement of THD+N at 1kHz and 5W into 4ohm load. This has very little value if we wanted to make a choice between amps with similar SINAD/THD+N. Measurement at one frequency covering static distortion only (1kHz) and one level is close to pointless, it only selects the worse pieces from the set.

 

[solderdude]

SINAD is great! Valid measurement. The problem is ASR's calling it a definitive measure of "guaranteed transparency." That's not science. That's belief.

Or it is your belief that it is not science.

 

[scott wurcer]

So, you agree that "120dB SINAD guarantees transparency" ? You agree that lower THD is always perceptively better? That's frankly all I'm here for. If you agree with either of these statements, the onus is on -you- to deliver the abundant peer-reviewed proof. Otherwise, it is opinion. And opinion is GREAT -- we all have opinions based on our personal experience. Just don't represent it as "audio science."

If you read what I said I have asked for more measurements. Don't put words in my mouth, IMO the full scale THD at best shows care in engineering and at worst a deliberate deviation from good engineering. I have never stated otherwise. You keep saying perceptively better, but I have run across folks in the music business and we have had two examples here where the unique mangling of the audio signal by poorly chosen transformers is preferred (and yes the simple A/B difference is a 10/10).

You don't see scientists wasting their time doing peer reviewed articles debunking things like the crack pot "Electric Universe" folks. In the engineering community at large the high open-loop gain gobs of feedback stuff is not worth any attention.

 

[Xulonn]

But, alas, this all returns to peer-reviewed AB testing. If there is a preponderance of Journal evidence to support a statement, I'm in. If there's not, it's not science, but opinion.

Ah, yes! Peer-review AB testing (in scientific journals) is the only valid foundation for judging. Please provide citations for the peer-reviewed journal papers that support your claims and describe your peer-reviewed testing methods.

I have seen a few comments by experts in this thread that appear to challenge your less than SOTA AB testing methods, methods that do not emply the highest standard - ABX testing. And yet you are adamant that those lower reliability "proprietary," non-published methods bu used for any testing "challenge" in which you would be willing to participate. You say "If there is a preponderance of Journal evidence to support a statement, I'm in. If there's not, it's not science, but opinion." That is an admission that your comments are just as much opinion as Amir's with respect to the validity of SINAD as a useful metric for estimating "engineering excellence" - and it illustrates a significant level of hypocrisy in your rantings.

Plus, you keep "embellishing" and misinterpreting the simple fact that in general, higher SINAD is better. Are you unable to comprehend the concept of general guidelines, and the appropriateness of imprecise technical generalities for consumer advice? I understand the debate about the usefulness and validity of a level of 116-120dB SINAD as a threshold of audibility. However, I see as your biased, non-scientific "opinion" about the lack of usefulness as a consumer guideline of SINAD rankings to be more less valid than the opinion of Amir (and me) that it is valid for low and semi-technical audio consumers. Although most of us non- and semi-experts seem to be able to see some usefulness in the SINAD bar chart and its categories, you seem to be utterly incapable of comprehending that it is a general classification system useful for establishing the range of engineering competence that a DAC belongs to, and not a precise scientific tool for establishing the audibility of distortion and noise of a particular DAC.

The most important aspect of the SINAD chart - a recommendation that many of us recommend to newbies - is that one should avoid the bottom tier (e.g., below 85dB SINAD), that a DAC in tier 4 of 5 is marginal, and that anything in the top three tiers is likely to be distortion and noise free in almost all HiFi audio systems.

And of course, most of us realize that some opinions hold more weight than others. Your opinion seems top be based no non-peer reviewed testing methods that support the saleable features and benefits of your products, while Amir's opinions appear to be based on a study of (and respect for) peer-reviewed science. Amir's expertise comes from years of experience where he occasionally pushed the limits of his abilities, and did not hesitate to question the results when appropriate. Being aware of the tricks the brain can play on one with respect to the perception of sonic differences and preferences under the umbrella of psycho acoustics is a critical aspect of being more objective in evaluating audibility of differences between components. And Amir demonstrates the humility of being proven wrong by science, and as a result is often quite adamant about what has been "proven" right or wrong beyond a reasonable doubt, and skeptical about that which violates the known limitations of human perception.

All opinions are not equal in weight, and some "opinions" are worth much more than others. I do consider the opinion of confirmed experts in a field over non-experts in that field.

Definition of scientific opinion: Scientific opinion means the opinion, within their respective fields, of competent engineers and scientists with experience in the area of discussion.

 

[amirm]

But, alas, this all returns to peer-reviewed AB testing. If there is a preponderance of Journal evidence to support a statement, I'm in. If there's not, it's not science, but opinion.

Well, you are not in because peer reviewed journal information were presented to you from JAES, all of which you ignored. Here they are again:

“Dynamic-Range Issues in the Modern Digital Audio Environment, ” Fielder, Louis D., JAES Volume 43 Issue 5 pp. 322-339; May 1995

That is exactly the type of analysis I have been showing and explaining to you.

“Noise: Methods for Estimating Detectability and Threshold, ” Stuart, J. Robert, JAES Volume 42 Issue 3 pp. 124-140; March 1994

Same analysis of detection threshold relative to Fletcher-Munson curves. This is the gold standard in proving audibility since it is broadly applicable. In contrast, you can test two opamps and people would say, "how about other opamps?" There is a difference between proof and data. I have been showing you proof. You keep asking for data which if presented, would totally invalidate your position anyway.

I suggest paying the $12 or whatever it costs to purchase these papers, and start on a path to understand the science here.

The above papers are old by the way because there is no interest in chasing the pink elephants you suggest. People fail to hear differences in gear at far lower levels than 120 dB. Yet you keep insisting on data to prove 120 dB SINAD is transparent? Give us a break please. We have heard your point of view but it doesn't hold any water. It is the most antiscience position to take and one that is chanted by people who completely ignore audio research across decades.

 

[signalpath]

Ah, yes! Peer-review AB testing (in scientific journals) is the only valid foundation for judging. Please provide citations for the peer-reviewed journal papers that support your claims and describe your peer-reviewed testing methods.

Peer-review is the only valid confirmation of truth and efficacy in science, in all branches of science. This is why homeopathic medicine is derided among medical science, but embraced by those easily duped.

Once again: I don't claim that my AB methods are accurate or even valid. Perhaps you missed the last half-dozen times I've stated that. I can only speak for our company, our 30 years of experience doing in-house AB testing, hundreds of hours of critical AB listening, and the results we've seen in-market. It works for us, but if we claimed "guaranteed scientific proof" of our methods and results, it would be laughed off the stage at an AES conference, simply because the sample size (2 or 3 guys in a lab) is invalid, among other reasons.

In the same vein, ASR is making THD claims that would, similarly, be laughed off the stage at AES. If the site was called "Audio Opinion Blog," then I wouldn't be wasting my breath. But ASR represents itself as a "science-based" site. I've made two simple suggestions that, if embraced, would put ASR back on track as a science-based destination. Right now, their THD beliefs are little different than homeopathy. I would love to be able to recommend ASR to my AES peers as a "real science based" audio site. Make it so, Mr. Data.

I'm still waiting for someone to come along with a standard collection of peer-reviewed science which shows that SINAD/THD and perceptive accuracy / transparency is necessarily correlated. When that information arrives, I'll immediately admit I was wrong, apologize for wasting everyone's time, and rest in the fact that there's always something new to learn (hint: such data is not extant).

Like yesterday. Someone told me that Australia has the largest worldwide population of wild CAMELS, like 1.5 million. I laughed it off. Then I looked it up. I didn't know. And I readily admit when I'm wrong. It's not embarrassing. It's encouraging and empowering and helps the entire human community. Try it!

 

[March Audio]

So, you agree that "120dB SINAD guarantees transparency" ? You agree that lower THD is always perceptively better? These are my only concerns with this ASR site.

Well they shouldn't be concerns because that is not what the ASR site says. This is purely your misinterpretation and, considering that it has been explained multiple times now, just a strawman you are using for the purpose of your argument.

It guarantees transparency of that metric. Below a certain level you are not going to perceive those specific issues. This does not equate to "this is the only measurement that has any audible relevance."

This also specifically contradicts your second point.

So can you please stop going round in circles?

 

[signalpath]

Well, you are not in because peer reviewed journal information were presented to you from JAES, all of which you ignored. Here they are again:

This is getting silly.

Why do you keep presenting dynamic range data, psychoacoustic data, JJ data, AES IC amp perception tests, etc.. in support of your claim that -120dB THD hardware performance guarantees transparency?

There's a connection missing between -120dB dynamic range perception limits (which we agree -- it's been settled, peer-reviewed audio science since the 1930s) and the broad ASR claim that -120dB THD/SINAD hardware product performance "guarantees transparency." That's a leap of religious faith, not accepted science.

I recognize that you've done "extensive AB testing" to support your claim. That's good. My company has done (and continues to do) "extensive testing" to improve the sonic quality of our products, and that's also good. But neither of us has a peer-reviewed standing to make seicntific claims based on our proprietary experiences.

If you want to promote the idea that -120dB SINAD guarantees transparency (as a scientific truth), I suggest you first arrange a major peer-reviewed ABX-qualified trial to test and validate your hypothesis. Get your most vocal critics intimately involved to assure that everyone's specific concerns are covered. I'll be first in line with a series of tests that we have found to be perceptual at the -120 SINAD level (micamps, etc.).

If ultimately a major test shows null positive to your postulate, hats off to you. And we all win. And if not. We all win.

v best regards.

 

[March Audio]

Ah, yes! Peer-review AB testing (in scientific journals) is the only valid foundation for judging. Please provide citations for the peer-reviewed journal papers that support your claims and describe your peer-reviewed testing methods.

I have seen a few comments by experts in this thread that appear to challenge your less than SOTA AB testing methods, methods that do not emply the highest standard - ABX testing. And yet you are adamant that those lower reliability "proprietary," non-published methods bu used for any testing "challenge" in which you would be willing to participate. You say "If there is a preponderance of Journal evidence to support a statement, I'm in. If there's not, it's not science, but opinion." That is an admission that your comments are just as much opinion as Amir's with respect to the validity of SINAD as a useful metric for estimating "engineering excellence" - and it illustrates a significant level of hypocrisy in your rantings.

Plus, you keep "embellishing" and misinterpreting the simple fact that in general, higher SINAD is better. Are you unable to comprehend the concept of general guidelines, and the appropriateness of imprecise technical generalities for consumer advice? I understand the debate about the usefulness and validity of a level of 116-120dB SINAD as a threshold of audibility. However, I see as your biased, non-scientific "opinion" about the lack of usefulness as a consumer guideline of SINAD rankings to be more less valid than the opinion of Amir (and me) that it is valid for low and semi-technical audio consumers. Although most of us non- and semi-experts seem to be able to see some usefulness in the SINAD bar chart and its categories, you seem to be utterly incapable of comprehending that it is a general classification system useful for establishing the range of engineering competence that a DAC belongs to, and not a precise scientific tool for establishing the audibility of distortion and noise of a particular DAC.

The most important aspect of the SINAD chart - a recommendation that many of us recommend to newbies - is that one should avoid the bottom tier (e.g., below 85dB SINAD), that a DAC in tier 4 of 5 is marginal, and that anything in the top three tiers is likely to be distortion and noise free in almost all HiFi audio systems.

And of course, most of us realize that some opinions hold more weight than others. Your opinion seems top be based no non-peer reviewed testing methods that support the saleable features and benefits of your products, while Amir's opinions appear to be based on a study of (and respect for) peer-reviewed science. Amir's expertise comes from years of experience where he occasionally pushed the limits of his abilities, and did not hesitate to question the results when appropriate. Being aware of the tricks the brain can play on one with respect to the perception of sonic differences and preferences under the umbrella of psycho acoustics is a critical aspect of being more objective in evaluating audibility of differences between components. And Amir demonstrates the humility of being proven wrong by science, and as a result is often quite adamant about what has been "proven" right or wrong beyond a reasonable doubt, and skeptical about that which violates the known limitations of human perception.

All opinions are not equal in weight, and some "opinions" are worth much more than others. I do consider the opinion of confirmed experts in a field over non-experts in that field.

Can I like twice?

 

[March Audio]

And until the results of that test are in -- do your readers a service and stop making non-scientific claims. Or rename the community to something w/o "science" in the headline.
regards.

...... Said the man who has presented no science to back up his position........

 

[scott wurcer]

Right now, their THD beliefs are little different than homeopathy.

No different than class A discrete amps with obsolete unobtainium JFET's are just perceptually better but the quality that makes them so is just simply un-measurable or quantifiable.

I have folks I consider friends doing the Pass Labs thing. I did an open-loop RIAA with the "blessed" FET's and shared it with anyone that cared because it's important to them. I know full well I can't hear any difference with a decent FET op-amp circuit, it's entertainment and fun with friends. Thorough my HD650's my LP's sound just as good as when I had a 1984 SOTA LP set up.

I guess the listening tests are off the table?

 

[RichB]

Perhaps someone can provide some gounding in SPL levels for common sounds.
I found these on this amplifier power estimator:

https://myhometheater.homestead.com/splcalculator.html

Is it correct that a rustling leaf is 10 dB?
I have measured my rooms noise level in the 40+ dB range, and yet I can hear a pin drop.

- Rich

 

[majingotan]

I mean you can still hear -10 to -20 dB from the noise floor of the room so it’s still possible for to hear a pin drop even at 50+ dB environments

 

[amirm]

This is getting silly.

Oh, it got silly from the start.

Why do you keep presenting dynamic range data, psychoacoustic data, JJ data, AES IC amp perception tests, etc.. in support of your claim that -120dB THD hardware performance guarantees transparency?

You really didn't understand that explanation? Then there is no hope for you on this topic.

There's a connection missing between -120dB dynamic range perception limits (which we agree -- it's been settled, peer-reviewed audio science since the 1930s) and the broad ASR claim that -120dB THD/SINAD hardware product performance "guarantees transparency." That's a leap of religious faith, not accepted science.

Do you just cut and paste this over and over again?

With respect to noise and distortion, that is what 120 dB SINAD does. It brings both distortion and noise below threshold of audibility if you listen at 120 dB SPL. If you listen at lower level, then you don't need that high of a SINAD.

Now, if you screw up the frequency response, run out of power and clip, etc., those are audible factors as well. In the case of DACs which is where you objection came from, they don't clip (but if they do it is noted in the review and SINAD takes a huge dive). Frequency response variation is next to non-existence and certainly so in our age bracket.

What is left is how much the DAC distorts or produces noise. Sum those in the form of SINAD and if they are below threshold of audibility, that is the end of the discussion. Protesting it over and over again because you sell devices that are not remotely this good doesn't amount to anything substantive.

I recognize that you've done "extensive AB testing" to support your claim. That's good. My company has done (and continues to do) "extensive testing" to improve the sonic quality of our products, and that's also good. But neither of us has a peer-reviewed standing to make seicntific claims based on our proprietary experiences.

I have told you nothing above that has to do with my extensive AB testing. I showed you peer reviewed science that is used to analyze audibility of a reproduction channel. Without reading or understanding of the topic, or even knowing what is an AB test or ABX, you keep protesting.

If you want to promote the idea that -120dB SINAD guarantees transparency (as a scientific truth), I suggest you first arrange a major peer-reviewed ABX-qualified trial to test and validate your hypothesis. Get your most vocal critics intimately involved to assure that everyone's specific concerns are covered. I'll be first in line with a series of tests that we have found to be perceptual at the -120 SINAD level (micamps, etc.).

If ultimately a major test shows null positive to your postulate, hats off to you. And we all win. And if not. We all win.

v best regards.

I am not the one making the fantastical claims. You are. Yet you have not shown us a bit of evidence. I keep asking you to post measurements. You have not. I asked you to show how you performed past tests with specifics of products, you have not post any. Come back with an argument that goes against the science as has been explained to you and then we see if we need to take action.

I have also offered $1,000 for you to perform a proper test and report back. You did not bite on that and are asking me again to do work? If I chased every wrong headed idea from subjectivist audio people like yourself, I would not have time to have lunch let alone do anything else.

I suggest instead of arguing with me, find someone who knows this topic, JJ would do, and get their feedback instead of wasting our time with repeated same lines. I know I am tired of this.

 

[LTig]

We're talking about electronic components - amplifiers. Would you agree that an amplifier with a constant gain within the audible band is transparent? Constant gain means just a fixed factor, no noise at all, no distortion at all:

Vout = gain * Vin​

If yes:

Unfortunately such an amplifier does not exist. Real amplifiers add unwanted artifacts to the signal:​

• linear distortion (FR)
• nonlinear distortion
• harmonic distortion
• nonharmonic distortion
• intermodulation distortion
• noise
• whatever ...

How high (or low) should be the levels for each kind of non wanted artifacts to be transparent?​

If no: what would be the technical characteristics of a transparent amplifier?

It would be helpful if @signalpath could answer these questions. It may shed some light into his thinking and may move the discussion into a more helpful direction.

 

[amirm]

I have measured my rooms noise level in the 40+ dB range, and yet I can hear a pin drop.

- Rich

That's because you don't want to use a single number SPL. It is a dumb metric until you convert it to a spectrum. Once there, you realize that in the mid frequencies where your hearing is very sensitive, actual SPL level of noise is quite a bit lower than 40 dB. See this paper I wrote which I have been referencing: https://www.audiosciencereview.com/forum/index.php?threads/dynamic-range-how-quiet-is-quiet.14/

 

[LTig]

In fact, we recently did a similar test for the most advanced micamp we've ever developed (premiering at NAMM in 2 weeks!). The test was to AB the sonic signature of various candidate input capacitors. At least two or three were rejected almost immediately. 10/10. A number were in the middle, OK but iffy. And two parts were clearly more transparent to the source, at least 8/10. We are using those qualified parts in the new product. (keep in mind, none of the capacitors under test impacted the baseline noise or THD performance of the signal path -- the source of detectable distortion is something other than THD -- which is another conversation that goes to root of why -120dB THD does not "guarantee transparency")

I don't dispute the results. Certain kind of coupling caps do distort; Douglas Self has published test results of his Ap in his book about small signal electronics. I'm wondering though that this is not measurable.

I'm even more puzzled when you write that opamps with 0.0001% THD sound worse than your discrete FET-amps with 0.02% THD. What's is the cause of the better sound you perceive with your micamps? If it's not distortion of the amps or the coupling caps - what is it, your something other? Can't you just tell it? Or is it a corporate secret you don't want to publish?

 

[March Audio]

Perhaps someone can provide some gounding in SPL levels for common sounds.
I found these on this amplifier power estimator:

https://myhometheater.homestead.com/splcalculator.html

View attachment 45077

Is it correct that a rustling leaf is 10 dB?
I have measured my rooms noise level in the 40+ dB range, and yet I can hear a pin drop.

- Rich

The issue here a simple SPL noise level covers the entire audible range (notional 20Hz to 20kHz). In simple terms its the sum of all the noise across that frequency range. However at any specific frequency the noise level is actually lower, so the rustling of a leaf and the frequencies that sound contains may be at a higher level than the background noise at those frequencies and therefore audible even if it is below the "overall" noise level.

Here is a measurement of the noise in my office. Its quite windy outside so levels will be high.

So whilst you can see in the box the overall level is 45dB(A), across the entire frequency range the levels are much lower than that. Add to this the fact that your hearing sensitivity changes with frequency then the single number SPL value isnt very useful.

 

[Xulonn]

Can I like twice?

LOL, Alan.

Unless the "industry expert who won't use his real name" "dude" can provide evidence that audio signal distortions below 120 dB can be heard, he is still "blowing smoke".

His beliefs seem to be based on experience-based opinions, but his non-rigorous, non-peer-reviewed AB testing methods may well be flawed, and producing invalid results - but he won't let anyone test for that. A big red flag for me was when he clearly and adamantly stated that his methods were not to be challenged and replaced by more rigorous, "science-approved" methods. Flailing about and whining about "non-science" by Amir here while refusing to use rigorous science himself truly is hypocritical.

These are the reasons I continue with my challenges to his flawed and baseless accusations. He admits to not using good scientific techniques himself, but yet demands that others conform to his beliefs. He is most definitely not a good arbiter of what is or is not "science".

 

[scott wurcer]

A big red flag for me was when he has clearly and adamantly stated that his methods were not to be challenged and replaced by more rigorous, "science-approved" methods. Flailing about and whining about "non-science" by Amir here while refusing to use rigorous science himself truly is hypocritical.

There is a long document on perceptual testing for small differences by one of the European standards committees but I lost the link to it. Maybe someone can find it. I suggest this be the process, warning it is very elaborate.

Three mates all with a financial interest in the results sitting together doing the testing is not part of the program.

I need to make the point that your products are probably wonderful that is not even remotely the issue. It's the claims that are. Some of the designers you reference believe in cable lifters, magic stones, quantum purifiers, freezing pictures of your amp, etc.

I apologize if I am wrong but I have a hard time believing that ALL of these snake oil things are completely absent from your lexicon.